Added Support vector machine code in C++

code/artificial_intelligence/src/support_vector_machine/support_vector_machine.cpp

@@ -0,0 +1,71 @@
+#include <iostream>
+#include <vector>
+using namespace std;
+
+class SVMModel {
+public:
+    vector<double> weights;
+    double bias;
+
+    SVMModel(int numFeatures) : weights(numFeatures, 0), bias(0) {}
+};
+
+class SVM {
+private:
+    static const double LEARNING_RATE;
+    static const int MAX_ITERATIONS;
+
+public:
+    static SVMModel train(const vector<vector<double>>& features, const vector<int>& labels) {
+        int numFeatures = features[0].size();
+        SVMModel model(numFeatures);
+
+        for (int iteration = 0; iteration < MAX_ITERATIONS; iteration++) {
+            for (size_t i = 0; i < features.size(); i++) {
+                const vector<double>& x = features[i];
+                int y = labels[i];
+
+                double prediction = predict(model.weights, x, model.bias);
+                double hingeLoss = max(0.0, 1 - y * prediction);
+
+                // Update weights and bias using gradient descent
+                if (hingeLoss > 0) {
+                    for (int j = 0; j < numFeatures; j++) {
+                        model.weights[j] -= LEARNING_RATE * (-y * x[j]);
+                    }
+                    model.bias -= LEARNING_RATE * (-y);
+                }
+            }
+        }
+
+        return model;
+    }
+
+    static double predict(const vector<double>& weights, const vector<double>& x, double bias) {
+        double result = bias;
+        for (size_t i = 0; i < weights.size(); i++) {
+            result += weights[i] * x[i];
+        }
+        return result;
+    }
+};
+
+const double SVM::LEARNING_RATE = 0.01;
+const int SVM::MAX_ITERATIONS = 1000;
+
+int main() {
+    vector<vector<double>> features = {{1, 2}, {2, 3}, {3, 4}};
+    // Add more training examples as needed
+
+    vector<int> labels = {1, 1, -1};
+    // Add corresponding labels for each training example
+
+    SVMModel model = SVM::train(features, labels);
+
+    // Test the trained model
+    vector<double> testExample = {4, 5};
+    double prediction = SVM::predict(model.weights, testExample, model.bias);
+    cout << "Prediction: " << prediction << endl;
+
+    return 0;
+}